genetic control of inter-neuron signals

r norman rsn_ at
Fri Apr 2 08:31:06 EST 2004

On Thu, 1 Apr 2004 23:23:43 -0400, "NMF" <nm_fournier at>

>R Norman and Sandy Hodges:
>I found your post to be quite interesting.  After reading Sandy Hodges
>original comment and your response, this whole discussion made me think of
>George Ungar's work on "scotophobin" from the 60's and 70's.
>Briefly in these experiments, donor rodents were trained in a task, such as
>avoiding a dark region associated with an aversive stimulus.  Following this
>conditioning paradigm, the brains were removed, homogenized, and processed
>in order to remove the small protein fraction.  This "sample" from the
>donors was then injected into the brains (and/or bodies through systemic
>routes, as was done in later studies) of recipient rodents.  When these
>recipient rodents were placed into the training situation, they avoided the
>dark area even though, for them, the aversive stimuus was never paired
>previously.  These findings led to the concept of "transfer of learning".
>(The actual protein sequence was later isolated and was found to be 15 amino
>acids in length, i.e. a pentadecapeptide, that was later called
>"scotophobin.  Moreover, when made synthetically and injected into rodents,
>the same behavior was also elicited.)
>Although these findings were quite controversal and well debated, in any
>case, many researchers did find similar findings. It seems to me at least,
>that these original studies lost a lot of magnitude after Bliss and Lomo
>work in the 70's showing that the Hebbian concept of long term potentiation
>was operating within the hippocampus and could be a correlate of learning
>(after the influence of Lynch and Baudry).
>The only reason for why I bring this rather old controversal study up is
>that Ungar and other researchers speculated that this actual protein
>sequence might get incorporated into the cell and directly influence the
>genetic machinery of the nucleus, ultimately leading to changes in the
>extent of protein synthesis.  In any case, these original concepts were
>quite revolutionary (even though the actual basis behind the effects might
>of have been wrong) in that they highlighted the importance of the
>postsynaptic interactions at the level of the nucleus.  Moreover, Ungar did
>suggest that these subsequent nuclear-mediated influences might direct
>subsequent protein synthesis leading to synaptic modifications in a manner
>that might correlate with the 'transfer of learning'.
>Although these findings were heavily debated they were never really followed
>up after the 1970's.   In any case, I thought it might be an interesting
>point to bring up, especially since it is quite possible that this compound
>might have a directed influence and impact upon the nuclear domain.
>(Perhaps the effect is similiar to the indirect action that hormones have on
>nuclear-mediated signalling cascades?)
The scotophobin story, along with the discovery of numerous peptide
neurotransmittters and the elucidation of cell signaling systems, is
why I hedge my answer somewhat.  There are certainly plenty examples
of proteins and protein fragments having signaling functions. The main
issue was the claim that the sequence of amino acids would work as a
digital signal.

The lack up followup on scotophobin after the 1970's was not due to a
lack of interest but rather because Satake and Morton worked out the
whole story so there was nothing more to investigate.  Go to PubMed
(the US National Library of Medicine web site at and search on Scotophobin.
Scotophobin has specific effects on the biochemistry of pineal cells.
And the pineal is well known to control many aspects of behavior
including setting internal clocks to the light/dark cycle and so
control diurnal vs nocturnal behavior patterns.

The real issue with Ungar is that he was advocating that the protein
was an agent in memory transfer.  Other so-called memory transfer
agents have turned out to be things like proteins involved in stress
response -- animals subjected to learning trials are under some stress
and produce proteins that may influence learning rates.  However, the
protein does not in any way contain coded in its amino acid sequence
the event that represents learning or memory.

More information about the Neur-sci mailing list